The first rules under the Artificial Intelligence Act (AI Act), which regulates AI based on risk, have come into effect in the European Union. The law bans certain use cases deemed to pose an “unacceptable risk,” such as manipulation through subconscious techniques and social scoring that could lead to discrimination.
According to TechCrunch, the EU’s executive body has published guidance explaining how developers can comply. These recommendations address prohibited AI applications, violations of which could result in significant fines—up to 7% of a company’s global turnover or €35 million, whichever is higher.
EU Guidance and Implementation Timeline
“The recommendations are designed to ensure uniform and effective application of the AI Act throughout the European Union,” the Commission stated. However, these recommendations are not legally binding, with final interpretation and enforcement resting with regulators and courts.
The Commission clarified that the document provides legal explanations and practical examples to help stakeholders understand and implement the law’s requirements. The recommendations are currently in draft form and will be finalized once translated into all official EU languages.
Although the AI Act came into force last year, its implementation is ongoing, notes NIX Solutions. Additional provisions will take effect in the coming months and years. Compliance oversight will be phased in, with EU member states required to designate monitoring bodies by August 2.
We’ll keep you updated as more details emerge about the AI Act’s enforcement.